{
  "best_metric": 0.08962889015674591,
  "best_model_checkpoint": "model_training/deepseek_prover_sft_no_err/checkpoints-random-09-07-09-00/checkpoint-450",
  "epoch": 3.6568527918781726,
  "eval_steps": 50,
  "global_step": 450,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04060913705583756,
      "grad_norm": 95.18575286865234,
      "learning_rate": 1e-05,
      "loss": 12.8904,
      "step": 5
    },
    {
      "epoch": 0.08121827411167512,
      "grad_norm": 60.47364044189453,
      "learning_rate": 2e-05,
      "loss": 9.6717,
      "step": 10
    },
    {
      "epoch": 0.1218274111675127,
      "grad_norm": 69.353759765625,
      "learning_rate": 3e-05,
      "loss": 4.9122,
      "step": 15
    },
    {
      "epoch": 0.16243654822335024,
      "grad_norm": 0.35820522904396057,
      "learning_rate": 4e-05,
      "loss": 0.6982,
      "step": 20
    },
    {
      "epoch": 0.20304568527918782,
      "grad_norm": 0.3532850742340088,
      "learning_rate": 5e-05,
      "loss": 0.1738,
      "step": 25
    },
    {
      "epoch": 0.2436548223350254,
      "grad_norm": 8.06154727935791,
      "learning_rate": 6e-05,
      "loss": 0.1606,
      "step": 30
    },
    {
      "epoch": 0.28426395939086296,
      "grad_norm": 0.33542144298553467,
      "learning_rate": 7e-05,
      "loss": 0.1632,
      "step": 35
    },
    {
      "epoch": 0.3248730964467005,
      "grad_norm": 0.2429077923297882,
      "learning_rate": 8e-05,
      "loss": 0.141,
      "step": 40
    },
    {
      "epoch": 0.36548223350253806,
      "grad_norm": 0.17707887291908264,
      "learning_rate": 9e-05,
      "loss": 0.132,
      "step": 45
    },
    {
      "epoch": 0.40609137055837563,
      "grad_norm": 0.5356423258781433,
      "learning_rate": 0.0001,
      "loss": 0.1388,
      "step": 50
    },
    {
      "epoch": 0.40609137055837563,
      "eval_loss": 0.13083037734031677,
      "eval_runtime": 577.2624,
      "eval_samples_per_second": 1.732,
      "eval_steps_per_second": 0.217,
      "step": 50
    },
    {
      "epoch": 0.4467005076142132,
      "grad_norm": 0.0570620633661747,
      "learning_rate": 9.996842891446092e-05,
      "loss": 0.1288,
      "step": 55
    },
    {
      "epoch": 0.4873096446700508,
      "grad_norm": 0.07483270764350891,
      "learning_rate": 9.987375552718133e-05,
      "loss": 0.1179,
      "step": 60
    },
    {
      "epoch": 0.5279187817258884,
      "grad_norm": 0.04857001453638077,
      "learning_rate": 9.971609939582557e-05,
      "loss": 0.1245,
      "step": 65
    },
    {
      "epoch": 0.5685279187817259,
      "grad_norm": 0.06141730025410652,
      "learning_rate": 9.9495659615402e-05,
      "loss": 0.121,
      "step": 70
    },
    {
      "epoch": 0.6091370558375635,
      "grad_norm": 0.07013677060604095,
      "learning_rate": 9.921271456683715e-05,
      "loss": 0.1082,
      "step": 75
    },
    {
      "epoch": 0.649746192893401,
      "grad_norm": 0.027853988111019135,
      "learning_rate": 9.886762156542428e-05,
      "loss": 0.1338,
      "step": 80
    },
    {
      "epoch": 0.6903553299492385,
      "grad_norm": 0.039150092750787735,
      "learning_rate": 9.846081640959007e-05,
      "loss": 0.1161,
      "step": 85
    },
    {
      "epoch": 0.7309644670050761,
      "grad_norm": 0.029366208240389824,
      "learning_rate": 9.79928128305494e-05,
      "loss": 0.1151,
      "step": 90
    },
    {
      "epoch": 0.7715736040609137,
      "grad_norm": 0.02762170508503914,
      "learning_rate": 9.746420184354334e-05,
      "loss": 0.1192,
      "step": 95
    },
    {
      "epoch": 0.8121827411167513,
      "grad_norm": 0.025405680760741234,
      "learning_rate": 9.687565100147939e-05,
      "loss": 0.1137,
      "step": 100
    },
    {
      "epoch": 0.8121827411167513,
      "eval_loss": 0.11551322042942047,
      "eval_runtime": 577.0742,
      "eval_samples_per_second": 1.733,
      "eval_steps_per_second": 0.217,
      "step": 100
    },
    {
      "epoch": 0.8527918781725888,
      "grad_norm": 0.02768566645681858,
      "learning_rate": 9.622790355191672e-05,
      "loss": 0.1067,
      "step": 105
    },
    {
      "epoch": 0.8934010152284264,
      "grad_norm": 0.028063081204891205,
      "learning_rate": 9.552177749846083e-05,
      "loss": 0.1103,
      "step": 110
    },
    {
      "epoch": 0.934010152284264,
      "grad_norm": 0.028980253264307976,
      "learning_rate": 9.475816456775313e-05,
      "loss": 0.1173,
      "step": 115
    },
    {
      "epoch": 0.9746192893401016,
      "grad_norm": 0.025315387174487114,
      "learning_rate": 9.393802908335977e-05,
      "loss": 0.1089,
      "step": 120
    },
    {
      "epoch": 1.015228426395939,
      "grad_norm": 0.026901153847575188,
      "learning_rate": 9.306240674798203e-05,
      "loss": 0.1137,
      "step": 125
    },
    {
      "epoch": 1.0558375634517767,
      "grad_norm": 0.025451259687542915,
      "learning_rate": 9.213240333552589e-05,
      "loss": 0.1054,
      "step": 130
    },
    {
      "epoch": 1.0964467005076142,
      "grad_norm": 0.0280075091868639,
      "learning_rate": 9.114919329468282e-05,
      "loss": 0.1299,
      "step": 135
    },
    {
      "epoch": 1.1370558375634519,
      "grad_norm": 0.029911836609244347,
      "learning_rate": 9.011401826578492e-05,
      "loss": 0.1009,
      "step": 140
    },
    {
      "epoch": 1.1776649746192893,
      "grad_norm": 0.029297418892383575,
      "learning_rate": 8.902818551280758e-05,
      "loss": 0.1016,
      "step": 145
    },
    {
      "epoch": 1.218274111675127,
      "grad_norm": 0.0310602355748415,
      "learning_rate": 8.789306627249985e-05,
      "loss": 0.0949,
      "step": 150
    },
    {
      "epoch": 1.218274111675127,
      "eval_loss": 0.10870195180177689,
      "eval_runtime": 577.2383,
      "eval_samples_per_second": 1.732,
      "eval_steps_per_second": 0.217,
      "step": 150
    },
    {
      "epoch": 1.2588832487309645,
      "grad_norm": 0.030549369752407074,
      "learning_rate": 8.6710094022727e-05,
      "loss": 0.0996,
      "step": 155
    },
    {
      "epoch": 1.299492385786802,
      "grad_norm": 0.033894069492816925,
      "learning_rate": 8.548076267221256e-05,
      "loss": 0.106,
      "step": 160
    },
    {
      "epoch": 1.3401015228426396,
      "grad_norm": 0.034311868250370026,
      "learning_rate": 8.420662467396547e-05,
      "loss": 0.1162,
      "step": 165
    },
    {
      "epoch": 1.380710659898477,
      "grad_norm": 0.040043774992227554,
      "learning_rate": 8.288928906477496e-05,
      "loss": 0.1058,
      "step": 170
    },
    {
      "epoch": 1.4213197969543148,
      "grad_norm": 0.0440133698284626,
      "learning_rate": 8.15304194332491e-05,
      "loss": 0.101,
      "step": 175
    },
    {
      "epoch": 1.4619289340101522,
      "grad_norm": 0.03563845902681351,
      "learning_rate": 8.013173181896283e-05,
      "loss": 0.0981,
      "step": 180
    },
    {
      "epoch": 1.50253807106599,
      "grad_norm": 0.039490845054388046,
      "learning_rate": 7.869499254536865e-05,
      "loss": 0.1002,
      "step": 185
    },
    {
      "epoch": 1.5431472081218274,
      "grad_norm": 0.044993579387664795,
      "learning_rate": 7.722201598920673e-05,
      "loss": 0.0897,
      "step": 190
    },
    {
      "epoch": 1.5837563451776648,
      "grad_norm": 0.04990144819021225,
      "learning_rate": 7.571466228923115e-05,
      "loss": 0.0996,
      "step": 195
    },
    {
      "epoch": 1.6243654822335025,
      "grad_norm": 0.04168887436389923,
      "learning_rate": 7.417483499714589e-05,
      "loss": 0.1103,
      "step": 200
    },
    {
      "epoch": 1.6243654822335025,
      "eval_loss": 0.10267892479896545,
      "eval_runtime": 577.34,
      "eval_samples_per_second": 1.732,
      "eval_steps_per_second": 0.217,
      "step": 200
    },
    {
      "epoch": 1.6649746192893402,
      "grad_norm": 0.04457447677850723,
      "learning_rate": 7.260447867371709e-05,
      "loss": 0.0978,
      "step": 205
    },
    {
      "epoch": 1.7055837563451777,
      "grad_norm": 0.0444829948246479,
      "learning_rate": 7.100557643309732e-05,
      "loss": 0.0963,
      "step": 210
    },
    {
      "epoch": 1.7461928934010151,
      "grad_norm": 0.04889954626560211,
      "learning_rate": 6.938014743846285e-05,
      "loss": 0.096,
      "step": 215
    },
    {
      "epoch": 1.7868020304568528,
      "grad_norm": 0.05787263438105583,
      "learning_rate": 6.773024435212678e-05,
      "loss": 0.098,
      "step": 220
    },
    {
      "epoch": 1.8274111675126905,
      "grad_norm": 0.047007594257593155,
      "learning_rate": 6.605795074334794e-05,
      "loss": 0.0944,
      "step": 225
    },
    {
      "epoch": 1.868020304568528,
      "grad_norm": 0.06226496770977974,
      "learning_rate": 6.436537845710903e-05,
      "loss": 0.0985,
      "step": 230
    },
    {
      "epoch": 1.9086294416243654,
      "grad_norm": 0.04688010737299919,
      "learning_rate": 6.265466494718732e-05,
      "loss": 0.1104,
      "step": 235
    },
    {
      "epoch": 1.9492385786802031,
      "grad_norm": 0.04706396162509918,
      "learning_rate": 6.092797057688495e-05,
      "loss": 0.0879,
      "step": 240
    },
    {
      "epoch": 1.9898477157360406,
      "grad_norm": 0.05411742627620697,
      "learning_rate": 5.918747589082853e-05,
      "loss": 0.0918,
      "step": 245
    },
    {
      "epoch": 2.030456852791878,
      "grad_norm": 0.046843238174915314,
      "learning_rate": 5.7435378861282585e-05,
      "loss": 0.0896,
      "step": 250
    },
    {
      "epoch": 2.030456852791878,
      "eval_loss": 0.09781802445650101,
      "eval_runtime": 577.4371,
      "eval_samples_per_second": 1.732,
      "eval_steps_per_second": 0.216,
      "step": 250
    },
    {
      "epoch": 2.0710659898477157,
      "grad_norm": 0.04870203882455826,
      "learning_rate": 5.567389211245485e-05,
      "loss": 0.0915,
      "step": 255
    },
    {
      "epoch": 2.1116751269035534,
      "grad_norm": 0.05594008415937424,
      "learning_rate": 5.390524012629824e-05,
      "loss": 0.0863,
      "step": 260
    },
    {
      "epoch": 2.152284263959391,
      "grad_norm": 0.06335064023733139,
      "learning_rate": 5.2131656433338506e-05,
      "loss": 0.0931,
      "step": 265
    },
    {
      "epoch": 2.1928934010152283,
      "grad_norm": 0.06813333183526993,
      "learning_rate": 5.035538079207488e-05,
      "loss": 0.1013,
      "step": 270
    },
    {
      "epoch": 2.233502538071066,
      "grad_norm": 0.054481592029333115,
      "learning_rate": 4.857865636051585e-05,
      "loss": 0.0889,
      "step": 275
    },
    {
      "epoch": 2.2741116751269037,
      "grad_norm": 0.06014859676361084,
      "learning_rate": 4.6803726863421725e-05,
      "loss": 0.0873,
      "step": 280
    },
    {
      "epoch": 2.314720812182741,
      "grad_norm": 0.09669474512338638,
      "learning_rate": 4.503283375883165e-05,
      "loss": 0.0825,
      "step": 285
    },
    {
      "epoch": 2.3553299492385786,
      "grad_norm": 0.0704524964094162,
      "learning_rate": 4.326821340745304e-05,
      "loss": 0.0862,
      "step": 290
    },
    {
      "epoch": 2.3959390862944163,
      "grad_norm": 0.05738632380962372,
      "learning_rate": 4.151209424848819e-05,
      "loss": 0.0835,
      "step": 295
    },
    {
      "epoch": 2.436548223350254,
      "grad_norm": 0.06273756176233292,
      "learning_rate": 3.976669398546451e-05,
      "loss": 0.0887,
      "step": 300
    },
    {
      "epoch": 2.436548223350254,
      "eval_loss": 0.09393562376499176,
      "eval_runtime": 576.8653,
      "eval_samples_per_second": 1.734,
      "eval_steps_per_second": 0.217,
      "step": 300
    },
    {
      "epoch": 2.4791878172588833,
      "grad_norm": 0.05062380060553551,
      "learning_rate": 3.803421678562213e-05,
      "loss": 0.0855,
      "step": 305
    },
    {
      "epoch": 2.519796954314721,
      "grad_norm": 0.06058161333203316,
      "learning_rate": 3.631685049639586e-05,
      "loss": 0.0866,
      "step": 310
    },
    {
      "epoch": 2.5604060913705586,
      "grad_norm": 0.05282078683376312,
      "learning_rate": 3.461676388250651e-05,
      "loss": 0.0836,
      "step": 315
    },
    {
      "epoch": 2.601015228426396,
      "grad_norm": 0.06284045428037643,
      "learning_rate": 3.293610388715048e-05,
      "loss": 0.0777,
      "step": 320
    },
    {
      "epoch": 2.6416243654822336,
      "grad_norm": 0.06628148257732391,
      "learning_rate": 3.127699292074683e-05,
      "loss": 0.095,
      "step": 325
    },
    {
      "epoch": 2.682233502538071,
      "grad_norm": 0.0586889423429966,
      "learning_rate": 2.964152618066508e-05,
      "loss": 0.0851,
      "step": 330
    },
    {
      "epoch": 2.7228426395939085,
      "grad_norm": 0.06476856768131256,
      "learning_rate": 2.8031769005319147e-05,
      "loss": 0.0876,
      "step": 335
    },
    {
      "epoch": 2.763451776649746,
      "grad_norm": 0.062451042234897614,
      "learning_rate": 2.6449754265968264e-05,
      "loss": 0.0836,
      "step": 340
    },
    {
      "epoch": 2.804060913705584,
      "grad_norm": 0.09133639186620712,
      "learning_rate": 2.4897479799518796e-05,
      "loss": 0.0795,
      "step": 345
    },
    {
      "epoch": 2.8446700507614215,
      "grad_norm": 0.06529180705547333,
      "learning_rate": 2.3376905885569182e-05,
      "loss": 0.0843,
      "step": 350
    },
    {
      "epoch": 2.8446700507614215,
      "eval_loss": 0.0914108008146286,
      "eval_runtime": 578.8025,
      "eval_samples_per_second": 1.728,
      "eval_steps_per_second": 0.216,
      "step": 350
    },
    {
      "epoch": 2.885279187817259,
      "grad_norm": 0.05800933018326759,
      "learning_rate": 2.1889952770883643e-05,
      "loss": 0.0806,
      "step": 355
    },
    {
      "epoch": 2.9258883248730965,
      "grad_norm": 0.06730043888092041,
      "learning_rate": 2.043849824442124e-05,
      "loss": 0.086,
      "step": 360
    },
    {
      "epoch": 2.966497461928934,
      "grad_norm": 0.05438319221138954,
      "learning_rate": 1.9024375265982384e-05,
      "loss": 0.0846,
      "step": 365
    },
    {
      "epoch": 3.0071065989847714,
      "grad_norm": 0.06161397323012352,
      "learning_rate": 1.764936965146773e-05,
      "loss": 0.0811,
      "step": 370
    },
    {
      "epoch": 3.047715736040609,
      "grad_norm": 0.0842302069067955,
      "learning_rate": 1.631521781767214e-05,
      "loss": 0.0736,
      "step": 375
    },
    {
      "epoch": 3.0883248730964468,
      "grad_norm": 0.06074338033795357,
      "learning_rate": 1.502360458946232e-05,
      "loss": 0.0801,
      "step": 380
    },
    {
      "epoch": 3.1289340101522845,
      "grad_norm": 0.06264619529247284,
      "learning_rate": 1.3776161072106702e-05,
      "loss": 0.0805,
      "step": 385
    },
    {
      "epoch": 3.1695431472081217,
      "grad_norm": 0.07572674006223679,
      "learning_rate": 1.257446259144494e-05,
      "loss": 0.0841,
      "step": 390
    },
    {
      "epoch": 3.2101522842639594,
      "grad_norm": 0.08485180884599686,
      "learning_rate": 1.1420026704498077e-05,
      "loss": 0.0835,
      "step": 395
    },
    {
      "epoch": 3.250761421319797,
      "grad_norm": 0.05600133538246155,
      "learning_rate": 1.031431128303153e-05,
      "loss": 0.0775,
      "step": 400
    },
    {
      "epoch": 3.250761421319797,
      "eval_loss": 0.09013215452432632,
      "eval_runtime": 578.5208,
      "eval_samples_per_second": 1.729,
      "eval_steps_per_second": 0.216,
      "step": 400
    },
    {
      "epoch": 3.2913705583756343,
      "grad_norm": 0.05972089618444443,
      "learning_rate": 9.258712672491415e-06,
      "loss": 0.0855,
      "step": 405
    },
    {
      "epoch": 3.331979695431472,
      "grad_norm": 0.07054904848337173,
      "learning_rate": 8.254563928638893e-06,
      "loss": 0.083,
      "step": 410
    },
    {
      "epoch": 3.3725888324873097,
      "grad_norm": 0.06772007048130035,
      "learning_rate": 7.3031331341093915e-06,
      "loss": 0.0852,
      "step": 415
    },
    {
      "epoch": 3.4131979695431474,
      "grad_norm": 0.06778612732887268,
      "learning_rate": 6.405621797022848e-06,
      "loss": 0.083,
      "step": 420
    },
    {
      "epoch": 3.4538071065989846,
      "grad_norm": 0.06333562731742859,
      "learning_rate": 5.563163333667099e-06,
      "loss": 0.0727,
      "step": 425
    },
    {
      "epoch": 3.4944162436548223,
      "grad_norm": 0.05736032873392105,
      "learning_rate": 4.776821637170526e-06,
      "loss": 0.0756,
      "step": 430
    },
    {
      "epoch": 3.53502538071066,
      "grad_norm": 0.0700037032365799,
      "learning_rate": 4.047589733971646e-06,
      "loss": 0.0805,
      "step": 435
    },
    {
      "epoch": 3.575634517766497,
      "grad_norm": 0.0645386204123497,
      "learning_rate": 3.376388529782215e-06,
      "loss": 0.0813,
      "step": 440
    },
    {
      "epoch": 3.616243654822335,
      "grad_norm": 0.05811276286840439,
      "learning_rate": 2.7640656466274782e-06,
      "loss": 0.0808,
      "step": 445
    },
    {
      "epoch": 3.6568527918781726,
      "grad_norm": 0.06826285272836685,
      "learning_rate": 2.2113943524323167e-06,
      "loss": 0.0807,
      "step": 450
    },
    {
      "epoch": 3.6568527918781726,
      "eval_loss": 0.08962889015674591,
      "eval_runtime": 577.8097,
      "eval_samples_per_second": 1.731,
      "eval_steps_per_second": 0.216,
      "step": 450
    }
  ],
  "logging_steps": 5,
  "max_steps": 492,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.8705510622704435e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}